Cyberstrategics

I often wonder why there is tension in the world of CEP about the complexity of the events and the required processing to identify complex events.  Then, about a month ago, I was helping a friend with her masters degree thesis.  She is about to graduate with her masters degree in international business and is working on her thesis project (a resort project).    We were reviewing some basic marketing principles about customer satisfaction and I noticed that there was a basic “Marketing 101′ formula and diagram that illustrated the degree of customer satisfaction based on the customer’s expectations versus the level of service provided by the supplier.

It was then that I realized why there is so much tension in the “CEP world”.

Folks who have been in event processing fields like network management (NMS) or security management for many years have a very high expectation for processing complex events.   Most of the network and security management platforms on the market have basic rule-based processing available “out of the box” and most of these platforms have had the capability to process events in near-real time for decades.   Adding a new “rules-based event processing platform” to the network and security management software mix does little to add any additional capability and certainly does not solve any nagging complex detection problems.

In other words, the expectations of the network and security management customers are very high.  Perhaps that is why we don’t see much progress or focus in this area by the so-called CEP vendors (nor do we here much support from the existing NMS vendors!).  Yes, a NMS or security management person might experiment with these “CEP engines”, but they will soon discovery that these “rule-based stream processors” (for a lack of a better term) do not solve the complex problems of false alarms versus detection accuracy.  NMS folks have rules engines.  NMS folks can process streaming events. What the NMS folks are missing are models for complex event analysis (causality), prediction models and ways to process complex events in the cloud of uncertainty that exists in all networks.

Perhaps things are different on the trading floors of capital markets?

I have no experience on the trading floor.  I come from the field of network and security management.  However, it seem that folks in capital markets are really happy with “rule-based stream processors”.  Perhaps this is because capital market (trading) customers have never had an “open engine” to process events before?  Perhaps they only had closed systems before?  Perhaps these gurus are overjoyed to be able to write their own rules against streaming market data?

Perhaps I understand why there is so much emotion and “lack of logic” in the arguments of folks associated with market data in capital markets.   Their expectations seem to be lower than folks in more detection-oriented areas like security and network management.   Folks in network and security management have been processing events with event processing platforms for over 20 years.   These customer expect more and therefore the self-described “CEP vendors” are not making much headway there.

However, in areas that may have been dominated by proprietary vendor solutions, perhaps order management (OMS) or order routing systems in capital markets,  may I assume there were no generic platforms to process those events 20 years ago?

Perhaps this is one of the main reasons why reasonable people in the field of CEP have such different opinions on their satisfaction with the so called “CEP vendors” and their software platforms.  Most of the passion in support of these vendors comes from the “feeds and speeds” side of capital markets.  In fact, we read very little support coming from other areas or domains.  The complimentary articles that come from analysts are almost always from folks working on the trading side of the financial services industry.

Obviously, it seems, the expectations of one group of people are different than those of folks in other industries.  Because the expectations are different, the degree of customer satisfaction is different. Hence, folks on the trading side of the spectrum should understand that folks in NMS and security management have been using open rule-based engines to process events for over 20 years.  We seek “something more”.  Our expectations are higher.

Likewise, folks on the NMS and security management side of the house, like myself, should understand, if it is indeed true, that folks on the FSI side of the house have not had similar event processing capabilities over the past 20 years and therefore these new “CEP engines” are very exciting to them.    Marketing 101?  Expectations are different so customer satisfaction is different?

Filed under: Complex Event Processing, Cyberstrategics, Event Cloud, Event Processing, Event Stream Processing, Financial Services, Systems Engineering, Use Cases